Botched upgrade to 44, no network in rescue terminal

Ask Fedora
1

Hi all,

Went ahead with trying to update my system yesterday to Fedora 44, everything seemed to be going fine but after leaving my computer for a while and coming back to it it just showed a black screen with a “_” for several hours. After rebooting I only noticed that 43 kernels are listed, so I guess something failed in the update.

I managed to get a rescue terminal after doing a chroot from a live USB to create a root password. My understanding is that the usual procedure in this case would be to do:

# dnf distro-sync --refresh --releasever=44 --setopt protected_packages=

However, I do not have access to the network so this fails right away. I tried following instructions to connect to a wireless network (no wired network in this laptop) from here:

But I run into this:

# nmcli general status
Error: Could not create NMClient object: Could not connect: No such file or directory.

Bit stuck at this point. Can always do a clean install but fixing the current one would be best. Any help would be greatly appreciated!

2

If you boot from a Live USB, I believe Anaconda has a recovery option.

3

If you use the Fedora KDE edition live USB, you can connect to a wireless network from the KDE GUI, then chroot into your install to do dnf distro-sync.

Hopefully there’s then no need to do anything with nmcli on the command line.

I assume it’s similar for most other desktop environments, but I’m not so familiar with them.

4

Just managed to get things up again, and indeed part of it involved doing the dnf distro-sync from the chroot in the live USB. However I ended up needing a few extra things.

  • I followed the post-install steps (Making sure you're not a bot!). This gave me the right kernel on boot, but boot still did not complete.
  • From journalctl logs I saw dbus-broker was failing to start due to empty configuration files. Decided to just apply the full reinstall from the chroot (sudo dnf reinstall $(rpm -qa --qf "%{NAME}\n") --skip-unavailable). Now but ended up stuck with the message “Failed to allocate manager object”.
  • Found this thread (Reboot stuck in "Failed to allocate manger object") which showed that error pops up due to messed up SELinux config when modifying things in chroot. Switching to permissive and setting up relabeling on boot as described in that thread made things work again!